Avoid counting characters and add explanatory comment to test

Author: varkor

src/liballoc/str.rs

@@ -380,7 +380,13 @@ impl str {
                reason = "return type may change to be an iterator",
                issue = "27791")]
     pub fn escape_debug(&self) -> String {
-        self.chars().enumerate().flat_map(|(i, c)| c.escape_debug_ext(i == 0)).collect()
+        let mut string = String::with_capacity(self.len());
+        let mut chars = self.chars();
+        if let Some(first) = chars.next() {
+            string.extend(first.escape_debug_ext(true))
+        }
+        string.extend(chars.flat_map(|c| c.escape_debug_ext(false)));
+        string
     }
 
     /// Escapes each char in `s` with [`char::escape_default`].

src/liballoc/tests/str.rs

@@ -989,6 +989,12 @@ fn test_escape_unicode() {
 
 #[test]
 fn test_escape_debug() {
+    // Note that there are subtleties with the number of backslashes
+    // on the left- and right-hand sides. In particular, Unicode code points
+    // are usually escaped with two backslashes on the right-hand side, as
+    // they are escaped. However, when the character is unescaped (e.g. for
+    // printable characters), only a single backslash appears (as the character
+    // itself appears in the debug string).
     assert_eq!("abc".escape_debug(), "abc");
     assert_eq!("a c".escape_debug(), "a c");
     assert_eq!("éèê".escape_debug(), "éèê");

src/libcore/unicode/unicode.py

@@ -21,7 +21,7 @@
 # - UnicodeData.txt
 #
 # Since this should not require frequent updates, we just store this
-# out-of-line and check the unicode.py file into git.
+# out-of-line and check the tables.rs file into git.
 
 import fileinput, re, os, sys, operator, math, datetime